小波多尺度变换对滚珠丝杠振动信号处理效果的研究

针对磨削加工中滚珠丝杠振动信号存在低信噪比及野值的特性,提出了基于正交小波多尺度分解的振动信号滤波方法。利用正交小波多尺度变化的基本原理,将振动信号分解到各个尺度上。在各个尺度上分别进行振动信号特性分析,之后通过正交小波重构得到高信噪比、去野值的振动信号。通过仿真试验及真实实验数据分析,验证了振动信号经过小波多尺度分解和滤波和重构后,不仅信噪比得到提高,并且振动信号的野值特性也可以得到很好地剔除。

Research on effect of WaVelet Multiseale Transform up on the vibration signal in Ball Screw

For solving the low signal noise ratio and outliers elimination property in vibration signal of ball screw grinding, the Vibration signal filter method based on orthogonal multiscale wavelet analysis is proposed. Using the orthogonal multiscale wavelet analysis theory, the vibration signal is decomposed to the various scales and its property is analyzed respectively, then the vibration signal of high signal noise ratio and removed outliers is obtained by orthogonal wavelet reconstruction. After wavelet multi-scale decomposition and filtering reconstructed, not only is the signal noise ratio of vibration signal increased, but also outliers elimination property is well removed by simulation and real data analysis.